Radiation Protection Dosimetry (2015), Vol. 164, No. 3, pp. 181 –186 Advance Access publication 10 September 2014
doi:10.1093/rpd/ncu272
COMPARISON OF THE NMIJ AND THE ARPANSA STANDARDS FOR ABSORBED DOSE TO WATER IN HIGH-ENERGY PHOTON BEAMS
*Corresponding author: [email protected] Received 15 April 2014; revised 25 July 2014; accepted 30 July 2014 The authors report the results of an indirect comparison of the standards of absorbed dose to water in high-energy photon beams from a clinical linac and 60Co radiation beam performed between the National Metrology Institute of Japan (NMIJ) and the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). Three ionisation chambers were calibrated by the NMIJ in April and June 2013 and by the ARPANSA in May 2013. The average ratios of the calibration coefficients for the three ionisation chambers obtained by the NMIJ to those obtained by the ARPANSA were 0.9994, 1.0040 and 1.0045 for 6-, 10- and 15-MV (18 MV at the ARPANSA) high-energy photon beams, respectively. The relative standard uncertainty of the value was 7.2` 3 1023. The ratio for 60Co radiation was 0.9986(66), which is consistent with the results published in the key comparison of BIPM.RI(I)-K4.
INTRODUCTION Absorbed dose to water in high-energy photon beams from a clinical linac is the quantity of interest for radiation cancer therapy. The National Metrology Institute of Japan (NMIJ) established the Japanese primary standard for absorbed dose to water in a 60 Co radiation beam based on a graphite calorimeter in 2009(1), and it disseminates the standard to secondary standard calibration laboratories. In Japan, the absorbed dose to water in high-energy photon beams is determined at each hospital by using the calibration coefficient for 60Co radiation and the beam quality correction factor, kQ, provided by the Japan Society of Medical Physics (JSMP)(2). In the kQ factor calculation, JSMP calculated the displacement correction factors by using the empirical formula of Wang and Rogers(3). The kQ factors by JSMP are 0.2– 0.4 % smaller than those listed in TRS-398(4). To reduce the uncertainty of the absorbed dose to water for highenergy photon beams, several National Metrology Institutes (NMIs) have established primary standards of absorbed dose to water in high-energy photon beams and provide a calibration service to secondary standard calibration laboratories or hospitals. These standards enable the determination of absorbed dose to water in clinical beams without using the calculated values of beam quality correction factor, which has a large standard uncertainty of 1.0 %(2). Likewise, the NMIJ has developed a graphite calorimeter as the Japanese primary standard of absorbed dose to water in high-energy photon beams from a clinical linac.
To confirm the dosimetric equivalence of the NMIJ standard of absorbed dose to water with the Bureau International des Poids et Mesures (BIPM) and other NMIs, the authors conducted a comparison between the NMIJ and the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). In the comparison, the NMIJ and the ARPANSA determined the calibration coefficients of three ionisation chambers for high-energy photon beams from the clinical linacs at the NMIJ and the ARPANSA, respectively. The BIPM.RI(I)-K6 key comparison of absorbed dose to water in high-energy photon beams with the ARPANSA was performed during the period September–October 2012(5, 6). Consequently, the comparison with the ARPANSA enables the NMIJ standard for absorbed dose to water to be compared with the BIPM and other participants of the BIPM.RI(I)-K6. Of particular interest in this comparison is that both the NMIJ and the ARPANSA standards are established using similar methods with graphite calorimeters. STANDARDS OF ABSORBED DOSE TO WATER AT THE NMIJ The NMIJ established the primary standard of absorbed dose to water in 60Co radiation using a graphite calorimeter and participated in the BIPM.RI(I)-K4 key comparison in 2009(7, 8). As a next step, the NMIJ developed a standard for high-energy photon beams from a clinical linac using the same graphite calorimeter.
# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
M. Shimizu1, *, Y. Morishita1, M. Kato1, T. Tanaka1, T. Kurosawa1, N. Takata1, N. Saito1, G. Ramanathan2, P. D. Harty2, C. Oliver2, T. Wright2,3 and D. J. Butler2 1 National Metrology Institute of Japan, AIST, Tsukuba, Japan 2 Australian Radiation Protection and Nuclear Safety Agency, Yallambie, Australia 3 School of Chemistry and Physics, University of Adelaide, Adelaide, Australia
M. SHIMIZU ET AL.
Figure 1 shows the linac (Elekta K. K., Precise) and the graphite calorimeter of the NMIJ. The graphite calorimeter is of Domen type(9). It has a graphite disc (called a core) on which two thermistor sensors and four thermistor heaters are mounted. The temperature rise of the core 4T (K) caused by photon beam irradiation is measured using the thermistor sensor. The absorbed dose to the core, DC (Gy), can be expressed as follows: DC ¼
ð1Þ
where Mcore (kg) and Ccore (J K21) denote the mass and heat capacity of the core, respectively. The heat capacity is measured from the reference electric heat UR (J) supplied to the core through the thermistor heaters: Ccore ¼
UR ; DTR
ð2Þ
where 4TR (K) is the temperature rise of the core due to the applied heat. The heat capacity was measured before and after each irradiation with the high-energy photon beams. The relative standard uncertainty of DC was 0.21 %.
where QW and QC denote the measured charge from a core-sized graphite cavity chamber placed in the water phantom and graphite calorimeter phantom at the respective reference points. The core-sized graphite cavity chamber is described in detail in (1). Monte Carlo (MC) indicates that the quantities in the brackets of MC ðD0W ; D0C ; Q0W ; Q0C Þ are obtained by using the MC calculation code EGS5(10) (M. Shimizu et al., in preparation). D0W and D0C are calculated absorbed dose to water and graphite, respectively. Q0W and Q0C are calculated charges from the cavity chamber in the water phantom and graphite calorimeter phantom, which are estimated by the calculated absorbed dose to the air in the cavity chamber. The uncertainty of the MC calculation, which includes the statistical uncertainties and the uncertainty of the incident electron beam profiles and those of the stopping power data of graphite and water, is 0.2 %. The combined relative standard uncertainty of CW,C is 0.29 %. The correction factor, kdef, is the heat defect correction of
Figure 1. NMIJ linac and graphite calorimeter.
182
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
Ccore DT ; Mcore
A factor, C W,C, to convert from the measured absorbed dose to the core to the absorbed dose to water is calculated using the following equation: QW D0 W =D0 C kdef kimp krn ; ð3Þ CW;C ¼ QC Q0 W =Q0 C MC
COMPARISON OF THE DOSE WITH NMIJ AND ARPANSA
graphite. This correction factor is treated as 1.0, and the uncertainty of kdef is 0.1 %(11). The authors took into account the effect of the impurities of the graphite calorimeter in the MC calculation; thus, the authors treated the impurity correction factor, kimp, as 1.0 and the uncertainty of this effect is 0.05 %. krn is the correction factor for the radial non-uniformity of the beam, and the uncertainty of krn is 0.1 %. The absorbed dose to water at the reference depth, DW, is then calculated by: DW ¼ DC CW;C :
ð4Þ
ND;W;NMIJ ¼
DW ; kTP kH kpol ks QNMIJ
ð5Þ
where kTP, kH, kpol and ks are the correction factors for temperature and pressure, humidity, polarity and ion recombination, respectively, QNMIJ is the charge measured by using the ionisation chamber and kpol and ks are calculated by using the procedure of TRS398(4). For the high-energy photon beams from the NMIJ’s linac, the polarity correction and the ion recombination correction are 1.0005 and 1.0037, respectively. For 60Co-gamma radiation, the measured charge is not corrected for the ion recombination. The combined uncertainty of the polarity correction and the ion recombination correction is 0.08 %. The relative standard uncertainty of the NMIJ calibration coefficients in all beams is estimated to be 0.4 %(1) (M. Shimizu et al., in preparation).
DW;ARP ¼ DC;ARP
krn ;
ð6Þ
MC
where krn is a correction for radial beam non-uniformity with respect to the central axis dose. krn is calculated over a volume identical to that of the calorimeter core. The ARPANSA primary standard of absorbed dose to water has been established for 60Co radiation and for 6-, 10- and 18-MV photon beams from the ARPANSA’s linac (Elekta Synergy). The ARPANSA has performed the key comparisons of BIPM.RI(I)-K4 (7, 15) and BIPM.RI(I)-K6(6). The reference conditions used by the ARPANSA are consistent with those recommended by the International Atomic Energy Agency in the Code of Practice TRS-398(4). The source to water surface distance (SSD) is 100 cm for 60Co and high-energy photon beams. The beam size is 10 cm` 10 cm at the surface of the water phantom for all beams. The reference water depth is 10 and 5 g cm22 for high-energy photon beams and 60Co radiation, respectively. The ARPANSA calibration coefficient of an ionisation chamber, ND,W,ARP, is calculated by the same method used by the NMIJ according to the following equation:
STANDARDS OF ABSORBED DOSE TO WATER AT THE ARPANSA
ND;W;ARP ¼
The ARPANSA primary standard for absorbed dose to water at both 60Co and high-energy photon beams is also based on graphite calorimetry. The calorimeter is of Domen design(9), and details are described in (12). The calorimeter has a core consisting of a
DW DC
DW;ARP ; kTP kH kpol ks QARP
ð7Þ
where QARP is the charge measured with the ionisation chamber in the ARPANSA reference conditions. For the high-energy photon beams from the ARPANSA’s linac, the polarity correction and the ion
183
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
The distance from the photon source to the reference point, i.e. the source to chamber distance (SCD), is 100 cm. The reference water depth along the beam axis in the water phantom is 10 g cm22 for high-energy photon beams. The beam size of the high-energy photon beams is 10` 10 cm at the reference point. For 60Co-gamma rays, the CW,C is calculated by using the cavity theory. The detail of the calculation and the NMIJ standard of absorbed dose to water in 60 Co radiation are described in (1). The reference water depth is 5 g cm22. The beam diameter of the 60Cogamma rays at the reference point (SCD ¼ 100 cm) is 11 cm. The calibration coefficient, ND,W,NMIJ, of an ionisation chamber at the reference point can be expressed as follows:
graphite disc of 20 mm in diameter and 2.75 mm in thickness. The core is surrounded by graphite bodies called the jacket, shield and medium. All the bodies are separated by evacuated gaps of 1 mm to reduce the heat loss and are fitted with sensing thermistors and heating elements. The core is fitted with two thermistors to sense the temperature rise and another as a heating element. Initial thermal equilibrium of all the bodies is maintained through continuous heating of the large medium by a heater control unit. The sensing thermistors are connected to a Wheatstone bridge circuit, and the output is processed by a lockin-amplifier. Measurements are converted from absorbed dose to the graphite core (DC,ARP) to absorbed dose to water (DW,ARP) using the MC simulation code BEAMnrc(13, 14). The graphite calorimeter and the water phantom are both modelled. In the calorimeter model, the absorbed dose to graphite in the calorimeter core is calculated. In the water phantom, the dose is calculated at a point at the reference depth over the same volume as the calorimeter core. The conversion factor is the ratio of calculated doses (DW/DC)MC. DW,ARP is calculated as follows:
M. SHIMIZU ET AL. Table 1. Calibration coefficients for the Exradin A12 and PTW 30013 ionisation chambers and the beam quality index (TPR20,10) of the NMIJ and the ARPANSA. ND,W,Q [mGy nC21]
NMIJ
Quality index TPR20,10
Exradin A12 SN-23331
PTW 30013 SN-1759
PTW 30013 SN-6740
60 Co 6 MV 10 MV 15 MV 60 Co 6 MV 10 MV 18 MV
0.569 0.679 0.729 0.758 0.569 0.673 0.734 0.777
47.65 47.24 47.00 46.69 47.80 47.29 46.72 46.18
53.58 52.80 52.40 51.98 53.58 52.91 52.19 51.48
53.43 52.65 52.25 51.84 53.49 52.72 52.00 51.28
recombination correction are 1.0005 and 1.0035, respectively. For 60Co-gamma radiation, the measured charge is not corrected for the ion recombination. The combined uncertainty of the polarity correction and the ion recombination correction is 0.08 %. The standard uncertainty of the ARPANSA calibration coefficients in all beams is 0.5 %.
COMPARISON PROCEDURE The NMIJ calibrated one Exradin A12 ionisation chamber and two PTW 30013 ionisation chambers for the 60Co radiation and for the 6-, 10-, and 15-MV photon beams at the NMIJ before and after the calibration by the ARPANSA. The ARPANSA calibrated these chambers for 60Co radiation and for the 6-, 10- and 18-MV high-energy photon beams at the ARPANSA. The reference calibration conditions were 20.08C and 101.325 kPa. The NMIJ reference field size (10` 10 cm at SCD ¼ 100 cm) at the chamber position is different from the ARPANSA reference field size (11` 11 cm at SCD ¼ 110 cm). No correction was made for the different field size and source– chamber distance. Because the NMIJ’s electrometer is not operated in a floating mode, a high voltage is applied to the ionisation chamber wall. Consequently, an insulator waterproof sleeve was used for the Exradin A12 whose chamber wall is made of conductive material. At the ARPANSA, all ionisation chambers were calibrated without a waterproof sleeve because a high voltage was applied to the central electrode of the chamber. In both the NMIJ and the ARPANSA measurements, central electrode positive was used. The sleeve effect is small, and the authors treat this effect as an uncertainty of 0.12 %(2).
Figure 2. Calibration coefficients of the Exradin A12 chamber for high-energy photon beams and 60Co-gamma rays. The solid and dashed lines show the quadratic fit function obtained for the ARPANSA and the NMIJ calibration coefficients, respectively.
RESULTS AND DISCUSSION Table 1 shows the values of the tissue–phantom ratio, TPR20,10, for all the high-energy photon beams of the NMIJ and the ARPANSA and the calibration coefficients for the ionisation chambers obtained by the NMIJ and the ARPANSA. The calibration coefficients are corrected for the polarity effect and ion recombination. For 60Co-gamma rays, the calibration coefficients of all chambers obtained by the NMIJ and the ARPANSA agree within the combined standard uncertainties given in Table 3. The greatest difference was 0.31 % for the A12 chamber, and the average ratio of NMIJ/ARPANSA was 0.9986(66). This result is consistent with the BIPM.R(I)-K4 key comparisons(8, 15) where the ratios of NMIJ/BIPM and ARPANSA/BIPM for 60Co were 0.9960(46) and 0.9973(53), respectively. The values in brackets are the
184
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
ARPANSA
Beam
COMPARISON OF THE DOSE WITH NMIJ AND ARPANSA
Table 2. Ratio of the calibration coefficients of the NMIJ to those of the ARPANSA at all beam qualities. ND;W;Q ðNMIJÞ=ND;W;Q ðARPANSA)
Beam 60 Co 6 MV
10 MV 15 MV 18 MV
Quality Exradin PTW PTW index A12 30013 30013 TPR20;10 SN-23331 SN-1759 SN-6740 Average 0.569 0.673 0.679 0.729 0.734 0.758 0.777
0.9969 1.0002 0.9998 1.0050 1.0040 1.0057 1.0082
1.0000 0.9995 0.9988 1.0031 1.0018 1.0033 1.0058
0.9988 1.0002 0.9995 1.0040 1.0027 1.0046 1.0072
0.9986 1.0000 0.9994 1.0040 1.0029 1.0045 1.0071
The averages of the values for the three ionisation chambers are also shown.
qualities (TPR20,10 ¼ 0.679, 0.729 and 0.758) were obtained from the fitted function. The values of TPR20,10 of the ARPANSA are 0.673, 0.734 and 0.777. The component of the relative standard uncertainty of the values of calibration coefficient due to the fitting is estimated to be 0.3 %(16, 17). Table 2 shows the ratios of the calibration coefficients between the NMIJ and the ARPANSA, which were obtained by interpolation or extrapolation to respective beam qualities shown in the table for all ionisation chambers. The relative standard uncertainty of the ratio for high-energy photon beams is 0.71 – 0.72 %, and the components of the uncertainty of the ratios for the high-energy photon beams and 60Cogamma rays are shown in Table 3. The ratios for two PTW 30013 chambers agreed with each other. However, the ratios for Exradin A12 chamber slightly differ from the ratios for PTW 30013 chambers. Both Exradin A12 and PTW 30013 are Farmer-type chambers, and the sensitive volumes of these chambers are similar to each other. Thus, the effect of the sensitive volume is negligibly small. The difference might be caused by the difference of the chamber materials and of the waterproof sleeve. In addition, the NMIJ’s electrometer applied a high voltage to the ionisation chamber wall, and an electric field is generated between the chamber wall and the waterproof sleeve. This electric field might increase the difference of Exradin A12 and PTW 30013. The NMIJ calibration coefficients for the 6-MV photon beam of the NMIJ and those of the ARPANSA (TPR20,10 ¼ 0.679 and 0.673) agree with the ARPANSA values. However, the NMIJ values for
Table 3. Standard uncertainty components of the ratio between the NMIJ and the ARPANSA calibration coefficients. Beam (A) PTW 30013 Co
60
High-energy photon beams
(B) Exradin A12 Co
60
High-energy photon beams
Quantity
Uncertainty (%)
NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Combined uncertainty NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Interpolation Combined uncertainty
0.4 0.5 0.64 0.4 0.5 0.3 0.71
NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Sleeve effect Combined uncertainty NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Interpolation Sleeve effect Combined uncertainty
0.4 0.5 0.12 0.66 0.4 0.5 0.3 0.12 0.72
Detailed uncertainty budgets can be found in (1, 6, 14 and M. Shimizu et al., in preparation).
185
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
combined relative standard uncertainties (k ¼ 1) for each value. In Figure 2, the calibration coefficients of the Exradin A12 chamber obtained by the NMIJ and the ARPANSA are plotted as a function of the tissue– phantom ratio, TPR20,10. To compare the values of calibration coefficients at the same TPR20,10, the values were obtained from the best-fit quadratic functions for the ARPANSA and the NMIJ calibration coefficients. The authors treated the TPR20,10 value of 60 Co-gamma rays as 0.569, which was measured at the ARPANSA. The best-fit lines are shown in Figure 2. The ARPANSA calibration coefficients that correspond to the TPR20,10 values of the NMIJ beam
M. SHIMIZU ET AL.
2.
3. 4.
5. 6.
CONCLUSION
7.
The authors performed a bilateral comparison of the standards of absorbed dose to water in high-energy photon beams between the NMIJ and the ARPANSA. Three ionisation chambers were calibrated for highenergy photon beams from two different clinical linacs at the NMIJ and the ARPANSA. The chambers were also calibrated for 60Co radiation beams. The values of the calibration coefficients at the same TPR20,10 were obtained by fitting quadratic functions to the respective results of calibration coefficients by the NMIJ and the ARPANSA. The mean values of the ratios between the NMIJ and the ARPANSA calibration coefficients of absorbed dose to water were 0.9986, 0.9994, 1.0040 and 1.0045 for the 60Co-gamma rays and the values of TPR20,10, which correspond to 6-, 10- and 15-MV beams at the NMIJ. The results agree within the relative standard uncertainty of 0.72 %. FUNDING
8.
9. 10. 11. 12.
13. 14.
This work is partially supported by the programme for young and women researchers of the University of Tokyo and used computational resources of the HPCI system provided by Kyushu University through the HPCI System Research Project (Project ID: hp120020). The calculation of the ARPANSA conversion factor was performed using the HPC facilities at eResearch South Australia.
15.
16.
REFERENCES 1. Morishita, Y., Kato, M., Takata, N., Kurosawa, T., Tanaka, T. and Saito, N. A standard for absorbed dose rate to water in a 60Co field using a graphite calorimeter
17.
186
at the national metrology institute of japan. Radiat. Prot. Dosim. 154(3), 331– 339 (2013). Japan Society of Medical Physics. Standard dosimetry of absorbed dose to water in external beam radiotherapy (standard dosimetry 12) Tusho Sangyo Kenkyusha K. K., (2012). Wang, L. L. W. and Rogers, D. W. O. The replacement correction factors for cylindrical chambers in high-energy photon beams. Phys. Med. Biol. 54, 1609– 1620 (2009). Andreo, P., Burns, D. T., Hohlfeld, K., Huq, M. S., Kanai, T., Laitano, F., Smythe, V. G. and Vynckier, S. Absorbed dose determination in external beam radiotherapy: an international code of practice for dosimetry based on standards of absorbed dose to water. Technical Report Series No. 398 International Atomic Energy Agency (2000). Key comparison BIPM.RI(I)-K6. BIPMkey Comparison Database, BIPM, http://kcdb.bipm.org/. Picard, S. et al. Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water of the ARPANSA, Australia and the BIPM in accelerator photon beams. Metrologia Technical Suppl. 51, 06006 (2014). Key comparison BIPM.RI(I)-K4. BIPM key Comparison Database, BIPM, http://kcdb.bipm.org/. Kessler, C, Allisy-Roberts, P. J., Morishita, Y, Kato, M, Takata, N, Kurosawa, T, Tanaka, T and Saito, N. Comparison of the standards for absorbed dose to water of the NMIJ and the BIPM for 60Co g-ray beams. Metrologia 48(1A), 06008 (2011). Domen, S. R. and Lamperti, P. J. A heat-loss-compensated calorimeter: theory, design, and performance. J. Res. NBS. A 78(5), 595 (1974). Hirayama, H., Namito, Y., Bielajew, A. F., Widerman, S. J. and Nelson, W. R. The EGS5 code system. SLACR-730 (2005) and KEK Report 2005– 8. Seuntjens, Jan and Duane, Simon. Photon absorbed dose standards. Metrologia 46(2), S39 (2009). Ramanathan, G., Harty, P., Wright, T., Lye, J. E., Butler, D. J., Webb, D. and Huntley, R. The Australian primary standard for absorbed dose to water (graphite calorimeter). ARPANSA Technical Report. 166, (2014). Walters, B., Kawrakow, I. and Rogers, D. W. O. BEAMnrc users manual. NRCC Report pages PIRS– 0509ArevK (2007). Lye, J. E., Butler, D. J., Franich, R. D., Harty, P. D., Oliver, C. P., Ramanathan, G., Webb, D. V. and Wright, T. Direct MC conversion of absorbed dose to graphite to absorbed dose to water for 60Co radiation. Radiat. Prot. Dosim. 155(1), 100–109 (2013). Kessler, C., Burns, D. T., Allisy, P. J., Butler, D., Lye, J. and Webb, D. Comparison of the standards for absorbed dose to water of the ARPANSA and the BIPM for 60Co g-radiation. Metrologia. 49(1A), 06009 (2012). Erazo, F. and Lallena, A. M. Calculation of beam quality correction factors for various thimble ionization chambers using the Monte Carlo code PENELOPE. Physica Medica. 29, 163– 170 (2013). Muir, B. R., McEwen, M. R. and Rogers, D. W. O. Measured and Monte Carlo calculated kQ factors: accuracy and comparison. Med. Phys. 38, 4600–4609 (2011).
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
10-MV (TPR20,10 ¼ 0.729 and 0.734) and 15-MV (TPR20,10 ¼ 0.758) photon beams are 0.4 % larger than their respective ARPANSA values. This small deviation may be caused by the difference between the methods used by the NMIJ and the ARPANSA for converting the graphite absorbed dose to the water absorbed dose. For 18-MV photon beam, the difference is ,0.82 %. The difference may be partly due to the large extrapolation of the NMIJ value. In the present comparison, the standards of absorbed dose to water of the NMIJ and the ARPANSA agree within 0.57 % for 60Co-gamma rays and 6-, 10- and 15-MV high-energy photon beams. The difference is less than the relative standard uncertainty of 0.72 %. Nevertheless, a trend that the discrepancy between the two results increases with the photon energy is evident and may warrant further investigation.
doi:10.1093/rpd/ncu272
COMPARISON OF THE NMIJ AND THE ARPANSA STANDARDS FOR ABSORBED DOSE TO WATER IN HIGH-ENERGY PHOTON BEAMS
*Corresponding author: [email protected] Received 15 April 2014; revised 25 July 2014; accepted 30 July 2014 The authors report the results of an indirect comparison of the standards of absorbed dose to water in high-energy photon beams from a clinical linac and 60Co radiation beam performed between the National Metrology Institute of Japan (NMIJ) and the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). Three ionisation chambers were calibrated by the NMIJ in April and June 2013 and by the ARPANSA in May 2013. The average ratios of the calibration coefficients for the three ionisation chambers obtained by the NMIJ to those obtained by the ARPANSA were 0.9994, 1.0040 and 1.0045 for 6-, 10- and 15-MV (18 MV at the ARPANSA) high-energy photon beams, respectively. The relative standard uncertainty of the value was 7.2` 3 1023. The ratio for 60Co radiation was 0.9986(66), which is consistent with the results published in the key comparison of BIPM.RI(I)-K4.
INTRODUCTION Absorbed dose to water in high-energy photon beams from a clinical linac is the quantity of interest for radiation cancer therapy. The National Metrology Institute of Japan (NMIJ) established the Japanese primary standard for absorbed dose to water in a 60 Co radiation beam based on a graphite calorimeter in 2009(1), and it disseminates the standard to secondary standard calibration laboratories. In Japan, the absorbed dose to water in high-energy photon beams is determined at each hospital by using the calibration coefficient for 60Co radiation and the beam quality correction factor, kQ, provided by the Japan Society of Medical Physics (JSMP)(2). In the kQ factor calculation, JSMP calculated the displacement correction factors by using the empirical formula of Wang and Rogers(3). The kQ factors by JSMP are 0.2– 0.4 % smaller than those listed in TRS-398(4). To reduce the uncertainty of the absorbed dose to water for highenergy photon beams, several National Metrology Institutes (NMIs) have established primary standards of absorbed dose to water in high-energy photon beams and provide a calibration service to secondary standard calibration laboratories or hospitals. These standards enable the determination of absorbed dose to water in clinical beams without using the calculated values of beam quality correction factor, which has a large standard uncertainty of 1.0 %(2). Likewise, the NMIJ has developed a graphite calorimeter as the Japanese primary standard of absorbed dose to water in high-energy photon beams from a clinical linac.
To confirm the dosimetric equivalence of the NMIJ standard of absorbed dose to water with the Bureau International des Poids et Mesures (BIPM) and other NMIs, the authors conducted a comparison between the NMIJ and the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). In the comparison, the NMIJ and the ARPANSA determined the calibration coefficients of three ionisation chambers for high-energy photon beams from the clinical linacs at the NMIJ and the ARPANSA, respectively. The BIPM.RI(I)-K6 key comparison of absorbed dose to water in high-energy photon beams with the ARPANSA was performed during the period September–October 2012(5, 6). Consequently, the comparison with the ARPANSA enables the NMIJ standard for absorbed dose to water to be compared with the BIPM and other participants of the BIPM.RI(I)-K6. Of particular interest in this comparison is that both the NMIJ and the ARPANSA standards are established using similar methods with graphite calorimeters. STANDARDS OF ABSORBED DOSE TO WATER AT THE NMIJ The NMIJ established the primary standard of absorbed dose to water in 60Co radiation using a graphite calorimeter and participated in the BIPM.RI(I)-K4 key comparison in 2009(7, 8). As a next step, the NMIJ developed a standard for high-energy photon beams from a clinical linac using the same graphite calorimeter.
# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
M. Shimizu1, *, Y. Morishita1, M. Kato1, T. Tanaka1, T. Kurosawa1, N. Takata1, N. Saito1, G. Ramanathan2, P. D. Harty2, C. Oliver2, T. Wright2,3 and D. J. Butler2 1 National Metrology Institute of Japan, AIST, Tsukuba, Japan 2 Australian Radiation Protection and Nuclear Safety Agency, Yallambie, Australia 3 School of Chemistry and Physics, University of Adelaide, Adelaide, Australia
M. SHIMIZU ET AL.
Figure 1 shows the linac (Elekta K. K., Precise) and the graphite calorimeter of the NMIJ. The graphite calorimeter is of Domen type(9). It has a graphite disc (called a core) on which two thermistor sensors and four thermistor heaters are mounted. The temperature rise of the core 4T (K) caused by photon beam irradiation is measured using the thermistor sensor. The absorbed dose to the core, DC (Gy), can be expressed as follows: DC ¼
ð1Þ
where Mcore (kg) and Ccore (J K21) denote the mass and heat capacity of the core, respectively. The heat capacity is measured from the reference electric heat UR (J) supplied to the core through the thermistor heaters: Ccore ¼
UR ; DTR
ð2Þ
where 4TR (K) is the temperature rise of the core due to the applied heat. The heat capacity was measured before and after each irradiation with the high-energy photon beams. The relative standard uncertainty of DC was 0.21 %.
where QW and QC denote the measured charge from a core-sized graphite cavity chamber placed in the water phantom and graphite calorimeter phantom at the respective reference points. The core-sized graphite cavity chamber is described in detail in (1). Monte Carlo (MC) indicates that the quantities in the brackets of MC ðD0W ; D0C ; Q0W ; Q0C Þ are obtained by using the MC calculation code EGS5(10) (M. Shimizu et al., in preparation). D0W and D0C are calculated absorbed dose to water and graphite, respectively. Q0W and Q0C are calculated charges from the cavity chamber in the water phantom and graphite calorimeter phantom, which are estimated by the calculated absorbed dose to the air in the cavity chamber. The uncertainty of the MC calculation, which includes the statistical uncertainties and the uncertainty of the incident electron beam profiles and those of the stopping power data of graphite and water, is 0.2 %. The combined relative standard uncertainty of CW,C is 0.29 %. The correction factor, kdef, is the heat defect correction of
Figure 1. NMIJ linac and graphite calorimeter.
182
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
Ccore DT ; Mcore
A factor, C W,C, to convert from the measured absorbed dose to the core to the absorbed dose to water is calculated using the following equation: QW D0 W =D0 C kdef kimp krn ; ð3Þ CW;C ¼ QC Q0 W =Q0 C MC
COMPARISON OF THE DOSE WITH NMIJ AND ARPANSA
graphite. This correction factor is treated as 1.0, and the uncertainty of kdef is 0.1 %(11). The authors took into account the effect of the impurities of the graphite calorimeter in the MC calculation; thus, the authors treated the impurity correction factor, kimp, as 1.0 and the uncertainty of this effect is 0.05 %. krn is the correction factor for the radial non-uniformity of the beam, and the uncertainty of krn is 0.1 %. The absorbed dose to water at the reference depth, DW, is then calculated by: DW ¼ DC CW;C :
ð4Þ
ND;W;NMIJ ¼
DW ; kTP kH kpol ks QNMIJ
ð5Þ
where kTP, kH, kpol and ks are the correction factors for temperature and pressure, humidity, polarity and ion recombination, respectively, QNMIJ is the charge measured by using the ionisation chamber and kpol and ks are calculated by using the procedure of TRS398(4). For the high-energy photon beams from the NMIJ’s linac, the polarity correction and the ion recombination correction are 1.0005 and 1.0037, respectively. For 60Co-gamma radiation, the measured charge is not corrected for the ion recombination. The combined uncertainty of the polarity correction and the ion recombination correction is 0.08 %. The relative standard uncertainty of the NMIJ calibration coefficients in all beams is estimated to be 0.4 %(1) (M. Shimizu et al., in preparation).
DW;ARP ¼ DC;ARP
krn ;
ð6Þ
MC
where krn is a correction for radial beam non-uniformity with respect to the central axis dose. krn is calculated over a volume identical to that of the calorimeter core. The ARPANSA primary standard of absorbed dose to water has been established for 60Co radiation and for 6-, 10- and 18-MV photon beams from the ARPANSA’s linac (Elekta Synergy). The ARPANSA has performed the key comparisons of BIPM.RI(I)-K4 (7, 15) and BIPM.RI(I)-K6(6). The reference conditions used by the ARPANSA are consistent with those recommended by the International Atomic Energy Agency in the Code of Practice TRS-398(4). The source to water surface distance (SSD) is 100 cm for 60Co and high-energy photon beams. The beam size is 10 cm` 10 cm at the surface of the water phantom for all beams. The reference water depth is 10 and 5 g cm22 for high-energy photon beams and 60Co radiation, respectively. The ARPANSA calibration coefficient of an ionisation chamber, ND,W,ARP, is calculated by the same method used by the NMIJ according to the following equation:
STANDARDS OF ABSORBED DOSE TO WATER AT THE ARPANSA
ND;W;ARP ¼
The ARPANSA primary standard for absorbed dose to water at both 60Co and high-energy photon beams is also based on graphite calorimetry. The calorimeter is of Domen design(9), and details are described in (12). The calorimeter has a core consisting of a
DW DC
DW;ARP ; kTP kH kpol ks QARP
ð7Þ
where QARP is the charge measured with the ionisation chamber in the ARPANSA reference conditions. For the high-energy photon beams from the ARPANSA’s linac, the polarity correction and the ion
183
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
The distance from the photon source to the reference point, i.e. the source to chamber distance (SCD), is 100 cm. The reference water depth along the beam axis in the water phantom is 10 g cm22 for high-energy photon beams. The beam size of the high-energy photon beams is 10` 10 cm at the reference point. For 60Co-gamma rays, the CW,C is calculated by using the cavity theory. The detail of the calculation and the NMIJ standard of absorbed dose to water in 60 Co radiation are described in (1). The reference water depth is 5 g cm22. The beam diameter of the 60Cogamma rays at the reference point (SCD ¼ 100 cm) is 11 cm. The calibration coefficient, ND,W,NMIJ, of an ionisation chamber at the reference point can be expressed as follows:
graphite disc of 20 mm in diameter and 2.75 mm in thickness. The core is surrounded by graphite bodies called the jacket, shield and medium. All the bodies are separated by evacuated gaps of 1 mm to reduce the heat loss and are fitted with sensing thermistors and heating elements. The core is fitted with two thermistors to sense the temperature rise and another as a heating element. Initial thermal equilibrium of all the bodies is maintained through continuous heating of the large medium by a heater control unit. The sensing thermistors are connected to a Wheatstone bridge circuit, and the output is processed by a lockin-amplifier. Measurements are converted from absorbed dose to the graphite core (DC,ARP) to absorbed dose to water (DW,ARP) using the MC simulation code BEAMnrc(13, 14). The graphite calorimeter and the water phantom are both modelled. In the calorimeter model, the absorbed dose to graphite in the calorimeter core is calculated. In the water phantom, the dose is calculated at a point at the reference depth over the same volume as the calorimeter core. The conversion factor is the ratio of calculated doses (DW/DC)MC. DW,ARP is calculated as follows:
M. SHIMIZU ET AL. Table 1. Calibration coefficients for the Exradin A12 and PTW 30013 ionisation chambers and the beam quality index (TPR20,10) of the NMIJ and the ARPANSA. ND,W,Q [mGy nC21]
NMIJ
Quality index TPR20,10
Exradin A12 SN-23331
PTW 30013 SN-1759
PTW 30013 SN-6740
60 Co 6 MV 10 MV 15 MV 60 Co 6 MV 10 MV 18 MV
0.569 0.679 0.729 0.758 0.569 0.673 0.734 0.777
47.65 47.24 47.00 46.69 47.80 47.29 46.72 46.18
53.58 52.80 52.40 51.98 53.58 52.91 52.19 51.48
53.43 52.65 52.25 51.84 53.49 52.72 52.00 51.28
recombination correction are 1.0005 and 1.0035, respectively. For 60Co-gamma radiation, the measured charge is not corrected for the ion recombination. The combined uncertainty of the polarity correction and the ion recombination correction is 0.08 %. The standard uncertainty of the ARPANSA calibration coefficients in all beams is 0.5 %.
COMPARISON PROCEDURE The NMIJ calibrated one Exradin A12 ionisation chamber and two PTW 30013 ionisation chambers for the 60Co radiation and for the 6-, 10-, and 15-MV photon beams at the NMIJ before and after the calibration by the ARPANSA. The ARPANSA calibrated these chambers for 60Co radiation and for the 6-, 10- and 18-MV high-energy photon beams at the ARPANSA. The reference calibration conditions were 20.08C and 101.325 kPa. The NMIJ reference field size (10` 10 cm at SCD ¼ 100 cm) at the chamber position is different from the ARPANSA reference field size (11` 11 cm at SCD ¼ 110 cm). No correction was made for the different field size and source– chamber distance. Because the NMIJ’s electrometer is not operated in a floating mode, a high voltage is applied to the ionisation chamber wall. Consequently, an insulator waterproof sleeve was used for the Exradin A12 whose chamber wall is made of conductive material. At the ARPANSA, all ionisation chambers were calibrated without a waterproof sleeve because a high voltage was applied to the central electrode of the chamber. In both the NMIJ and the ARPANSA measurements, central electrode positive was used. The sleeve effect is small, and the authors treat this effect as an uncertainty of 0.12 %(2).
Figure 2. Calibration coefficients of the Exradin A12 chamber for high-energy photon beams and 60Co-gamma rays. The solid and dashed lines show the quadratic fit function obtained for the ARPANSA and the NMIJ calibration coefficients, respectively.
RESULTS AND DISCUSSION Table 1 shows the values of the tissue–phantom ratio, TPR20,10, for all the high-energy photon beams of the NMIJ and the ARPANSA and the calibration coefficients for the ionisation chambers obtained by the NMIJ and the ARPANSA. The calibration coefficients are corrected for the polarity effect and ion recombination. For 60Co-gamma rays, the calibration coefficients of all chambers obtained by the NMIJ and the ARPANSA agree within the combined standard uncertainties given in Table 3. The greatest difference was 0.31 % for the A12 chamber, and the average ratio of NMIJ/ARPANSA was 0.9986(66). This result is consistent with the BIPM.R(I)-K4 key comparisons(8, 15) where the ratios of NMIJ/BIPM and ARPANSA/BIPM for 60Co were 0.9960(46) and 0.9973(53), respectively. The values in brackets are the
184
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
ARPANSA
Beam
COMPARISON OF THE DOSE WITH NMIJ AND ARPANSA
Table 2. Ratio of the calibration coefficients of the NMIJ to those of the ARPANSA at all beam qualities. ND;W;Q ðNMIJÞ=ND;W;Q ðARPANSA)
Beam 60 Co 6 MV
10 MV 15 MV 18 MV
Quality Exradin PTW PTW index A12 30013 30013 TPR20;10 SN-23331 SN-1759 SN-6740 Average 0.569 0.673 0.679 0.729 0.734 0.758 0.777
0.9969 1.0002 0.9998 1.0050 1.0040 1.0057 1.0082
1.0000 0.9995 0.9988 1.0031 1.0018 1.0033 1.0058
0.9988 1.0002 0.9995 1.0040 1.0027 1.0046 1.0072
0.9986 1.0000 0.9994 1.0040 1.0029 1.0045 1.0071
The averages of the values for the three ionisation chambers are also shown.
qualities (TPR20,10 ¼ 0.679, 0.729 and 0.758) were obtained from the fitted function. The values of TPR20,10 of the ARPANSA are 0.673, 0.734 and 0.777. The component of the relative standard uncertainty of the values of calibration coefficient due to the fitting is estimated to be 0.3 %(16, 17). Table 2 shows the ratios of the calibration coefficients between the NMIJ and the ARPANSA, which were obtained by interpolation or extrapolation to respective beam qualities shown in the table for all ionisation chambers. The relative standard uncertainty of the ratio for high-energy photon beams is 0.71 – 0.72 %, and the components of the uncertainty of the ratios for the high-energy photon beams and 60Cogamma rays are shown in Table 3. The ratios for two PTW 30013 chambers agreed with each other. However, the ratios for Exradin A12 chamber slightly differ from the ratios for PTW 30013 chambers. Both Exradin A12 and PTW 30013 are Farmer-type chambers, and the sensitive volumes of these chambers are similar to each other. Thus, the effect of the sensitive volume is negligibly small. The difference might be caused by the difference of the chamber materials and of the waterproof sleeve. In addition, the NMIJ’s electrometer applied a high voltage to the ionisation chamber wall, and an electric field is generated between the chamber wall and the waterproof sleeve. This electric field might increase the difference of Exradin A12 and PTW 30013. The NMIJ calibration coefficients for the 6-MV photon beam of the NMIJ and those of the ARPANSA (TPR20,10 ¼ 0.679 and 0.673) agree with the ARPANSA values. However, the NMIJ values for
Table 3. Standard uncertainty components of the ratio between the NMIJ and the ARPANSA calibration coefficients. Beam (A) PTW 30013 Co
60
High-energy photon beams
(B) Exradin A12 Co
60
High-energy photon beams
Quantity
Uncertainty (%)
NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Combined uncertainty NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Interpolation Combined uncertainty
0.4 0.5 0.64 0.4 0.5 0.3 0.71
NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Sleeve effect Combined uncertainty NMIJ calibration coefficient: ND,W,NMIJ ARPANSA calibration coefficient: ND,W,ARP Interpolation Sleeve effect Combined uncertainty
0.4 0.5 0.12 0.66 0.4 0.5 0.3 0.12 0.72
Detailed uncertainty budgets can be found in (1, 6, 14 and M. Shimizu et al., in preparation).
185
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
combined relative standard uncertainties (k ¼ 1) for each value. In Figure 2, the calibration coefficients of the Exradin A12 chamber obtained by the NMIJ and the ARPANSA are plotted as a function of the tissue– phantom ratio, TPR20,10. To compare the values of calibration coefficients at the same TPR20,10, the values were obtained from the best-fit quadratic functions for the ARPANSA and the NMIJ calibration coefficients. The authors treated the TPR20,10 value of 60 Co-gamma rays as 0.569, which was measured at the ARPANSA. The best-fit lines are shown in Figure 2. The ARPANSA calibration coefficients that correspond to the TPR20,10 values of the NMIJ beam
M. SHIMIZU ET AL.
2.
3. 4.
5. 6.
CONCLUSION
7.
The authors performed a bilateral comparison of the standards of absorbed dose to water in high-energy photon beams between the NMIJ and the ARPANSA. Three ionisation chambers were calibrated for highenergy photon beams from two different clinical linacs at the NMIJ and the ARPANSA. The chambers were also calibrated for 60Co radiation beams. The values of the calibration coefficients at the same TPR20,10 were obtained by fitting quadratic functions to the respective results of calibration coefficients by the NMIJ and the ARPANSA. The mean values of the ratios between the NMIJ and the ARPANSA calibration coefficients of absorbed dose to water were 0.9986, 0.9994, 1.0040 and 1.0045 for the 60Co-gamma rays and the values of TPR20,10, which correspond to 6-, 10- and 15-MV beams at the NMIJ. The results agree within the relative standard uncertainty of 0.72 %. FUNDING
8.
9. 10. 11. 12.
13. 14.
This work is partially supported by the programme for young and women researchers of the University of Tokyo and used computational resources of the HPCI system provided by Kyushu University through the HPCI System Research Project (Project ID: hp120020). The calculation of the ARPANSA conversion factor was performed using the HPC facilities at eResearch South Australia.
15.
16.
REFERENCES 1. Morishita, Y., Kato, M., Takata, N., Kurosawa, T., Tanaka, T. and Saito, N. A standard for absorbed dose rate to water in a 60Co field using a graphite calorimeter
17.
186
at the national metrology institute of japan. Radiat. Prot. Dosim. 154(3), 331– 339 (2013). Japan Society of Medical Physics. Standard dosimetry of absorbed dose to water in external beam radiotherapy (standard dosimetry 12) Tusho Sangyo Kenkyusha K. K., (2012). Wang, L. L. W. and Rogers, D. W. O. The replacement correction factors for cylindrical chambers in high-energy photon beams. Phys. Med. Biol. 54, 1609– 1620 (2009). Andreo, P., Burns, D. T., Hohlfeld, K., Huq, M. S., Kanai, T., Laitano, F., Smythe, V. G. and Vynckier, S. Absorbed dose determination in external beam radiotherapy: an international code of practice for dosimetry based on standards of absorbed dose to water. Technical Report Series No. 398 International Atomic Energy Agency (2000). Key comparison BIPM.RI(I)-K6. BIPMkey Comparison Database, BIPM, http://kcdb.bipm.org/. Picard, S. et al. Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water of the ARPANSA, Australia and the BIPM in accelerator photon beams. Metrologia Technical Suppl. 51, 06006 (2014). Key comparison BIPM.RI(I)-K4. BIPM key Comparison Database, BIPM, http://kcdb.bipm.org/. Kessler, C, Allisy-Roberts, P. J., Morishita, Y, Kato, M, Takata, N, Kurosawa, T, Tanaka, T and Saito, N. Comparison of the standards for absorbed dose to water of the NMIJ and the BIPM for 60Co g-ray beams. Metrologia 48(1A), 06008 (2011). Domen, S. R. and Lamperti, P. J. A heat-loss-compensated calorimeter: theory, design, and performance. J. Res. NBS. A 78(5), 595 (1974). Hirayama, H., Namito, Y., Bielajew, A. F., Widerman, S. J. and Nelson, W. R. The EGS5 code system. SLACR-730 (2005) and KEK Report 2005– 8. Seuntjens, Jan and Duane, Simon. Photon absorbed dose standards. Metrologia 46(2), S39 (2009). Ramanathan, G., Harty, P., Wright, T., Lye, J. E., Butler, D. J., Webb, D. and Huntley, R. The Australian primary standard for absorbed dose to water (graphite calorimeter). ARPANSA Technical Report. 166, (2014). Walters, B., Kawrakow, I. and Rogers, D. W. O. BEAMnrc users manual. NRCC Report pages PIRS– 0509ArevK (2007). Lye, J. E., Butler, D. J., Franich, R. D., Harty, P. D., Oliver, C. P., Ramanathan, G., Webb, D. V. and Wright, T. Direct MC conversion of absorbed dose to graphite to absorbed dose to water for 60Co radiation. Radiat. Prot. Dosim. 155(1), 100–109 (2013). Kessler, C., Burns, D. T., Allisy, P. J., Butler, D., Lye, J. and Webb, D. Comparison of the standards for absorbed dose to water of the ARPANSA and the BIPM for 60Co g-radiation. Metrologia. 49(1A), 06009 (2012). Erazo, F. and Lallena, A. M. Calculation of beam quality correction factors for various thimble ionization chambers using the Monte Carlo code PENELOPE. Physica Medica. 29, 163– 170 (2013). Muir, B. R., McEwen, M. R. and Rogers, D. W. O. Measured and Monte Carlo calculated kQ factors: accuracy and comparison. Med. Phys. 38, 4600–4609 (2011).
Downloaded from http://rpd.oxfordjournals.org/ at Nanyang Technological University on May 22, 2015
10-MV (TPR20,10 ¼ 0.729 and 0.734) and 15-MV (TPR20,10 ¼ 0.758) photon beams are 0.4 % larger than their respective ARPANSA values. This small deviation may be caused by the difference between the methods used by the NMIJ and the ARPANSA for converting the graphite absorbed dose to the water absorbed dose. For 18-MV photon beam, the difference is ,0.82 %. The difference may be partly due to the large extrapolation of the NMIJ value. In the present comparison, the standards of absorbed dose to water of the NMIJ and the ARPANSA agree within 0.57 % for 60Co-gamma rays and 6-, 10- and 15-MV high-energy photon beams. The difference is less than the relative standard uncertainty of 0.72 %. Nevertheless, a trend that the discrepancy between the two results increases with the photon energy is evident and may warrant further investigation.
